Table 4.
Comparison with other methods. Most of the methods classify X-rays only into pneumonia, COVID, and normal, our method further classifies COVID-19 into 3 categories
| Method | Type of images | Number of cases | Method used | Accuracy (%) |
|---|---|---|---|---|
| Ioannis et al. [14] | Chest X-ray | 1428 | VGG-19 | 93.48 |
| Wang et al. [13] | Chest X-ray | 13645 | COVID-Net | 92.4 |
| Sethy et al. [16] | Chest X-ray | 50 | ResNet50+SVM | 95.38 |
| Hemdan et al. [12] | Chest X-ray | 50 | COVIDX-Net | 90 |
| Narin et al. [15] | Chest X-ray | 100 | Deep CNN | 98 |
| Ying et al. [25] | Chest CT | 1485 | DRE-Net | 86 |
| Zheng et al. [9] | Chest CT | 542 | Unet+3D Deep Network | 90.8 |
| Xu et al. [11] | Chest CT | 618 | ResNet+ Location Attention | 86.7 |
| Proposed method | Chest X-ray | 22 | VGG-16 DenseNet-161 ResNet-18 | 95.9 98.9 76 |